Skip to main content

We've Moved!

Product Documentation has moved to docs.hitachivantara.com
Hitachi Vantara Knowledge

About LUN Manager, logical units (LUs), and host groups

LUN Manager operations

Your storage system can be connected to open-system server hosts of different platforms (for example, UNIX servers and PC servers). To configure your storage system for operation with open-system hosts, use LUN Manager to configure logical volumes and ports.

One of the important tasks when configuring logical volumes is to define I/O paths from hosts to logical volumes. When paths are defined, the hosts can send commands and data to the logical volumes and can receive data from the logical volumes.

After the system begins operating, you might need to modify the system configuration. For example, if hosts or drives are added, you will need to add new I/O paths. You can modify the system configuration using LUN Manager when the system is running. You do not need to restart the system after modifying the system configuration.

Host groups and LU path configuration (Fibre Channel)

After open-system hosts and the storage system are physically connected by cables, hubs, and so on, you can establish I/O paths between the hosts and the logical volumes. This defines which host can access which logical volume. Logical volumes that can be accessed by open-system hosts are referred to as logical units (LUs). The paths between the open-system hosts and the LUs are referred to as LU paths.

Before defining LU paths, you must classify server hosts by host groups. For example, if Linux hosts and Windows hosts are connected to the storage system, you must create one host group for the Linux hosts and another host group for the Windows hosts. Then, you must register the host bus adapters of the Linux hosts in the Linux host group, and you must also register the host bus adapters of the Windows hosts in the Windows host group.

A host group can contain only hosts that are connected to the same port. For example, if two Windows hosts are connected to port 1A and three Windows hosts are connected to port 1B, you cannot register all five Windows hosts in one host group. You must register the two Windows hosts on port 1A in one host group, and then register the three Windows hosts on port 1B in another host group.

After server hosts are classified into host groups, you associate the host groups with logical volumes. The following figure illustrates LU path configuration in a Fibre Channel environment. The figure shows host group hg-lnx associated with three logical volumes (00:00:00, 00:00:01, and 00:00:02). LU paths are defined between the two hosts in the hg-lnx group and the three logical volumes.

GUID-C2A78058-D89D-4F9B-801A-3A8F67CC5128-low.gif

You can define paths between a single server host and multiple LUs. The figure shows that each of the two hosts in the host group hg-lnx can access the three LUs.

You can also define paths between multiple server hosts and a single LU. The figure shows that the LU identified by the LDKC:CU:LDEV number 00:00:00 is accessible from the two hosts that belong to the hg-lnx host group.

The figure also shows that the LUs associated with the hg-lnx host group are addressed by numbers 0000 to 0002. The address number of an LU is referred to as a LUN (logical unit number). When software manipulates LUs, the software use LUNs to specify the LUs to be manipulated.

You can add, change, and delete LU paths when the system is in operation. For example, if new disks or server hosts are added to your storage system, you can add new LU paths. If an existing server host is to be replaced, you can delete the LU paths that correspond to the host before replacing the host. You do not need to restart the system when you add, change, or delete LU paths.

If a hardware failure (for example, CHA failure) occurs, there is a chance that some LU paths are disabled and some I/O operations are stopped. To avoid such a situation, you can define alternate LU paths; if one LU path fails, the alternate path takes over the host I/O.

Workflow for configuring logical units (Fibre Channel)

Use the following workflow to configure LUs in a Fibre Channel environment.

  1. Configure the ports.
  2. Configure the hosts.
  3. Configure the LU paths.
  4. Enable LUN security.
  5. Set Fibre Channel authentication.
  6. Manage the hosts.

Rules, restrictions, and guidelines for managing LUs

Rules
  • In a Fibre Channel environment, up to 2,048 LU paths can be defined for one host group, and up to 2,048 LU paths can be defined for one port.
  • In an iSCSI environment, up to 2,048 LU paths can be defined for one iSCSI target, and up to 2,048 LU paths can be defined for one port.
  • Up to 255 host groups can be created for one Fibre Channel port.
  • Up to 255 iSCSI targets can be created for one iSCSI port.
  • For an LDEV with the ALU attribute, you can define the LU path to only one host group.
  • For an LDEV with the ALU attribute, you can define the LU path to only one iSCSI target.
Restrictions
  • You cannot define an LU path to the following types of volumes:
    • Journal volumes
    • Pool volumes
    • External volumes with the data direct mapping attribute
    • LDEVs created from an accelerated compression-enabled parity group
    • Deduplication system data volumes
  • When defining LU paths, you must not use Command Control Interface and Hitachi Device Manager - Storage Navigator at the same time. If both programs are used simultaneously, operations might not be performed in the expected order, and the storage configuration might be defined incorrectly.
  • If an LDEV of the ALU attribute is binding to LDEVs with the SLU attribute, the LU path cannot be removed.
  • To define an LU path between a port and an LDEV that has the T10 PI attribute enabled, the port must have T10 PI mode enabled.
Guidelines
  • Queue depth: To ensure smooth processing at the ports and best average performance, the recommended queue depth setting (max tag count) for VSP Gx00 models and VSP Fx00 models is 1,024 per port and 32 per LDEV. Other queue depth settings, higher or lower than these recommended values, can provide improved performance for certain workload conditions.
    Caution Higher queue depth settings (greater than 1,024 per port) can impact host response times or cause failures such as job abend. Caution must be exercised in modifying the recommended queue depth settings.
  • Hitachi multipathing best practice is Single-Initiator Single-Targets configuration in which each HBA has only one path to the same LU. For example, you can provide four paths to each LU if you have four HBAs. For details, contact customer support.
  • If you attempt to apply many settings in the LUN Manager windows, the SVP might be unable to continue processing. Therefore, you should make no more than approximately 1,000 settings. Note that many settings are likely to be made when defining alternate paths, even though only two commands are required for defining alternate paths.
  • Do not perform the following when host I/O is in progress and hosts are in reserved status (mounted):
    • Remove LU paths
    • Disable LUN security on a port
    • Change the data transfer speed for Fibre Channel ports
    • Change AL-PAs or loop IDs
    • Change settings of fabric switches
    • Change the topology
    • Change the host modes
    • Remove host groups
    • Remove iSCSI targets
    • Setting command devices
  • When a port has a path defined for an LDEV with the GAD reserve attribute and you need to configure the port (delete a host group, set a command device, or change the host mode, topology, AL-PA, or transfer speed), the operation might fail if you configure more than one port at a time. For ports that have a path defined for an LDEV with the GAD reserve attribute, perform these operations on one port at a time.